Jump to content
8 posts in this topic

Recommended Posts

According to this article ( http://arstechnica.com/reviews/hardware/macbookpro.ars/6 ) the performance of the hacked version 10.4.4 is barely better than an old G4. I'm still wondering why maxxuss is using the old kernel back from the early dev kits... Maxxuss is there any way to fix this? Maybe you can investigate better in hacking the new kernel than replacing it with the old one. See results below:

 

bench5.png

Link to comment
https://www.insanelymac.com/forum/topic/11016-maxuss-please-fix-performance/
Share on other sites

i'm cautious about this. it says they "borrowed a friends laptop with OSX 10.4.4 installed on it". I wonder how optimised it was? Did it have all the correct drivers working? Was the gfx card working with QI/CI? And why is the Dell only on 10.4.4 and the others on 10.4.5?

 

I would like to see these benchmarks run again, this time against a fully optimised x86 install.

 

saying that though, it would be nice if we could get updated kernels. Maybe Maxxuss is waiting for 10.4.6 first (which should be released sometime in the next month)

He is not using a kernel from the Dev DVD's as far as I know it is the kernel from 10.4.4, I do not see anything saying it is the old one. I think that it was tested on an older machine with 256mb of ram propably to make people think it isnt worth it. the 10.4.5 release has the kernel from 10.4.4.

According to this article ( http://arstechnica.com/reviews/hardware/macbookpro.ars/6 ) the performance of the hacked version 10.4.4 is barely better than an old G4. I'm still wondering why maxxuss is using the old kernel back from the early dev kits... Maxxuss is there any way to fix this? Maybe you can investigate better in hacking the new kernel than replacing it with the old one. See results below:

 

bench5.png

 

huh that dell notebook is just thrash with a thrash cpu and that comparision is retarded anyway who idiot would compare a dualcore pentium m against a mobile p4

Hmmm, I have a P4 3.0GHz HT and 1gig of ram with an x1600 pro.... and my xbench score is 81.32, it only falls down in the thread test (no surprise there, only one core) and a mysterious problem where my uncached write of 4K blocks is only 1.21MB/sec *shrug*

 

something screwy there....

 

Results	81.32	
System Info		
	Xbench Version		1.2
	System Version		10.4.5 (8G1454)
	Physical RAM		1024 MB
	Model		ADP2,1
	Drive Type		ST380817AS
CPU Test	69.62	
	GCD Loop	98.89	5.21 Mops/sec
	Floating Point Basic	74.48	1.77 Gflop/sec
	vecLib FFT	51.38	1.70 Gflop/sec
	Floating Point Library	69.19	12.05 Mops/sec
Thread Test	109.30	
	Computation	89.40	1.81 Mops/sec, 4 threads
	Lock Contention	140.60	6.05 Mlocks/sec, 4 threads
Memory Test	63.64	
	System	70.76	
		Allocate	80.85	296.90 Kalloc/sec
		Fill	79.47	3864.10 MB/sec
		Copy	57.32	1184.00 MB/sec
	Stream	57.82	
		Copy	55.72	1150.83 MB/sec
		Scale	56.64	1170.23 MB/sec
		Add	59.98	1277.75 MB/sec
		Triad	59.14	1265.15 MB/sec
Quartz Graphics Test	92.80	
	Line	97.63	6.50 Klines/sec [50% alpha]
	Rectangle	86.08	25.70 Krects/sec [50% alpha]
	Circle	86.88	7.08 Kcircles/sec [50% alpha]
	Bezier	98.29	2.48 Kbeziers/sec [50% alpha]
	Text	96.78	6.05 Kchars/sec
OpenGL Graphics Test	137.83	
	Spinning Squares	137.83	174.85 frames/sec
User Interface Test	120.24	
	Elements	120.24	551.86 refresh/sec
Disk Test	48.76	
	Sequential	91.95	
		Uncached Write	91.04	55.89 MB/sec [4K blocks]
		Uncached Write	87.90	49.74 MB/sec [256K blocks]
		Uncached Read	91.42	26.76 MB/sec [4K blocks]
		Uncached Read	98.00	49.25 MB/sec [256K blocks]
	Random	33.18	
		Uncached Write	11.47	1.21 MB/sec [4K blocks]
		Uncached Write	69.02	22.10 MB/sec [256K blocks]
		Uncached Read	92.49	0.66 MB/sec [4K blocks]
		Uncached Read	124.12	23.03 MB/sec [256K blocks]

According to this article ( http://arstechnica.com/reviews/hardware/macbookpro.ars/6 ) the performance of the hacked version 10.4.4 is barely better than an old G4. I'm still wondering why maxxuss is using the old kernel back from the early dev kits... Maxxuss is there any way to fix this? Maybe you can investigate better in hacking the new kernel than replacing it with the old one. See results below:

 

He is using the kernel from 10.4.4. Just read his site. That laptop ars used must suck and be improperly configured, as stated before most people with 3.0Ghz P4s are getting ~80 in Xbench.

You should really read more of the review and this forum before asking for "fix performance"
Overall scores for the three machines were about what was expected, with the MacBook Pro standing at 53.71, the AlBook G4 at 45.73, and the Dell P4 at 22.96—a score no doubt affected by the exceptionally poor performance in the User Interface test. Without that score factored in, the Dell performed pretty well comparatively. One can only imagine what Mac OS X would be like on it if it were actually supported!
This can be fixed by disabling Beam Synchronizing with Quartz Debug:
I also ran XBench with Beam Synchronizing disabled on the MacBook Pro (which requires you to install the Developer Tools) and got a score of 161.15 in the User Interface test, a significant increase from the previous score of 27.52.
You can find a thread from this forum here: http://forum.osx86project.org/index.php?s=...indpost&p=65433
×
×
  • Create New...